Imagine a farmer who has a large yard and knows what impact the weather can have on his crop. Not only sun and rain can influence the growth of harvest, but also frost events that happen during the spring and fall can also have major impact and even destroy a full harvest within hours.  An automated frost prediction alert solution can be developed where actual measured data from an orchard, historical climate data and weather forecast data are combined, analyzed and used for providing growers early frost warnings (up to 5 days in advance) so that they can take appropriate countermeasures and save their crop and reaching their yield targets. With the capabilities of R and our package geoSpectRal, a model like this can be built.

What is geoSpectRal
It is an R package providing a new S4 class for R that stores spectral, temporal and spatial attributes of measurement data. Spectral data is usually obtained through radiometric measurements of the electromagnetic energy at various wavelengths. The package also offers methods for accessing and manipulating the spectral (and non-spectral) data. geoSpectRal was initially created to answer needs encountered in bio-optical oceanography, remote sensing, environmental and earth sciences. However anyone using spatial/temporal/spectral data of any kind can possibly make use of tools and methods provided by the package. The following S4 classes:

  • Spectra class (stores spatial/temporal/spectral aspects of data),
  • SpcHeader class (stores metadata in an R list object),
  • SpcList class (makes a collection of Spectra objects in an R list),

and basic data access and manipulation methods for importing, accessing and subsetting data, converting to/from R data.frames, analyzing, plotting, exporting scientific data are provided. Once spectral data is imported into a geoSpectral object, the statistical and data processing power of R is available for various kinds of scientific analyses.

GeoSpectRal is now CRAN certified, this means that our product geoSpectRal can be used on 13 platforms from different kind of flavours (from Linux to Windows to MacOS and many others).

Basics for geoSpectRal
It was build on top of another R package called spacetime, which, itself, was built on the packages rgdal (built on sp) and xts (built on zoo). rgdal and sp provide spatial attributes and methods. xts and zoo provide temporal attributes and methods time series analysis methods. spacetime facilitates spatio-temporal characterization of data, standardizing the way space and time attributes are stored in the variable. geoSpectral adds the spectral attributes and a simple metadata structure on top of the STIDF object provided by spacetime.

The package is issued with a GPLv3 license.

More information